• The forum software that supports hummy.tv has been upgraded to XenForo 2.3!

    Please bear with us as we continue to tweak things, and feel free to post any questions, issues or suggestions in the upgrade thread.

2TB HDD Corruption: Repair options?

S

SunnysideUK

Member
I have a Seagate 2TB Pipeline drive that got partially corrupted.
It had about 1.9TB of data on it but now it says it has about 1.8TB free (3.5% used).
I'm currently copying off files to another drive before attempting a "repair" as the "repair" may just clear its file table leaving any accessible files lost.
Q. What is the better repair tool: CF's fixdisk or tools within VMware 6.1 virtual Linux Mint 21.2? (I'm currently using this to do the copy).
 
Last edited:
SunnysideUK said:
I'm currently copying off files to another drive before attempting a "repair" as the "repair" may just clear its file table leaving any accessible files lost.
Q. What is the better repair tool: CF's fixdisk or tools within VMware 6.1 virtual Linux Mint 21.2? (I'm currently using this to do the copy).
Click to expand...
I would say fix-disk.
 
SunnysideUK said:
I'm currently copying off files to another drive before attempting a "repair" as the "repair" may just clear its file table leaving any accessible files lost.
Click to expand...
No need. The CF fixdisk option (from the Maintenance Mode menu) is as non-destructive as they come. Anything you lose will not have been available for copy.
 
It'll be a lot quicker running e2fsck on a real PC than on the Humax, but you have to do everything manually, including check the SMART stats. first.
I've thought before about making a PC friendly version of fix-disk, but never got round to actually tweaking it.
 
Likewise, but motivation failed because basically once you've got to the "disk out" stage you might as well recover any precious content to known good storage (ddrescue, even) and then wipe the suspect disk to clear any pending bad sectors. Mileage could vary for users with fewer scrap disks lying around.
 
I'm not having any success with fixdisk.
The volume remains in read-only state (webif: Filesystem /mnt/hd2 on /dev/sdb2 is read-only).
I've ran fixdisk several times. The VFD shows nt_stor:1x7 (presumed 7.?%) for some time, however telnet gives:

Error storing directory block information (inode=16389429, block=0, num=8105821): Memory allocation failed
e2fsck: aborted

NOTE:
I wanted to put this in a scrollable window but I've forgotten how.
I think its to do with [ ] /code /code etc. I can't work out how [ ] above work.
I've been going around in circles trying to find the wiki/help on this as I've seen it before but now can't find it. frustrating!
*update. found forum forum help bottom right of this page!
n.b. This site was inaccessible for a while recently.

[/code]
Code: 
More snippets of my log here:
07:35 06-12-2024
fixdisk....    note:
VFD shows: nt_stor:1x7 (7.?%), for longish time, just gone 1x8 at 07:05 06-12-2024,
and looking at telnet the screen is wizing by,,,

09:16 06-12-2024
sample when not streaming past screen to current end (FLD says 1-8. not moving. cannot hear head seeking):
           
Inode 15226912 has compression flag set on filesystem without

Inode 15226912 has INDEX_FL flag set but is not a directory.
Clear HTree index? yes

Inode 15226912, i_size is 12348808552957995606, should be 0.

Inode 15226912, i_blocks is 1694457532, should be 0.  Fix? yes

Inode 15226713 has a bad extended attribute block 93383692.  C

Inode 15226713, i_size is 12163236089735633282, should be 0.

Inode 15226713, i_blocks is 2278704183, should be 0.  Fix? yes

Inode 15226829 has compression flag set on filesystem without

Inode 15226829, i_size is 3770137135492121708, should be 0.  F

Inode 15226829, i_blocks is 528108893, should be 0.  Fix? yes

Inode 15226532 has illegal block(s).  Clear? yes

Illegal block #0 (2039850373) in inode 15226532.  CLEARED.
Illegal block #1 (2226084979) in inode 15226532.  CLEARED.
Illegal block #2 (1649778014) in inode 15226532.  CLEARED.
Illegal block #3 (4176595832) in inode 15226532.  CLEARED.
Illegal block #4 (4183432233) in inode 15226532.  CLEARED.
Illegal block #6 (3981156793) in inode 15226532.  CLEARED.
Illegal block #7 (4107099909) in inode 15226532.  CLEARED.
Illegal block #9 (4066161883) in inode 15226532.  CLEARED.
------------------------------------------------------------------------
09:00 06-12-2024    hard reboot - webif hdd still in read only mode.

12:24 06-12-2024    telnet fixdisk
14:00 06-12-2024    VHD: shows FixDisk Done,  ends with:

Illegal block #7 (2105020307) in inode 16389429.  CLEARED.
Illegal indirect block (4101034609) in inode 16389429.  CLEARED.
Illegal double indirect block (2521894659) in inode 16389429.  CLEARED.
Illegal block #11536398 (3937464914) in inode 16389429.  CLEARED.
Error storing directory block information (inode=16389429, block=0, num=8105821): Memory allocation failed
e2fsck: aborted

I was unable to exit to the menu as usual,
terminated telnet, reconnected and was offered connect.
now shows screen data flying past!
14:42 06-12-2024    ok, x to restart....


hmx_int_stor: ***** FILE SYSTEM WAS MODIFIED *****

hmx_int_stor: ***** FILE SYSTEM WAS MODIFIED *****
Fri Dec  6 13:58:29 GMT 2024
Insufficient memory to remove swap. Very large fix-disk log?
------------------------------------------------------------------------
14:00 06-12-2024    reboot
webif Msg:    Filesystem /mnt/hd2 on /dev/sdb2 is read-only
[code/]
----------------------------------
I'm thinking of formatting the video partition and manually copying back may well be quicker!
Could this be so?
If it is what's the format command to use please.
and does it only effect the recording, not settings?
 
Last edited:
Sometimes the filesystem structures are screwed in ways that e2fsck can't repair, or not on a machine like the setup box. Maybe some value is corrupted but not so much that it's obviously wrong; then the checker may try to create some data structure sized according to the corrupt value, and then "Memory allocation failed". Or this may be a bug in the checker, or fixdisk hasn't properly set up the paging space (aka swapfile).

At this point you have a choice:
1. (disk in) Copy off precious content from the problem disk and reformat the disk in situ, after which its fixdisk should succeed (if there are any bad sectors left) and whatever was readable of the precious content can be restored
2. (disk out) Swap in a new disk, copying over precious content from the problem disk if wanted; then erase and reformat the problem disk for some other purpose, perhaps as USB add-on storage for the Humax.

If the filesystem is badly screwed, there may be valid readable files that can't be reached from the root directory, because a directory entry in the file's path has been lost. e2fsck would try to find those for you and either restore a valid path or link recovered files into the /lost+found directory. Maybe a bigger machine will enable that to work (without the crash above), either on the problem disk or after _ddrescue_ing it to another disk or to an image file on another disk; otherwise missing content may be found online, or maybe TestDisk will find it.
 
SunnysideUK said:
I think its to do with [ ] /code /code etc. I can't work out how [ ] above work.
Click to expand...
You seem to have got there in the end, but you've left some dross behind. From Newbies' Guide to the Forum:
Black Hole said:
Code Listings and Terminal Output

We prefer that instances of code dumps or screen captures from command line sessions etc are embedded within a code box. This has several advantages: the fact that it is in a code box identifies it as separate from the discussion in the post; the font in the code box is fixed-pitch, so tabulated data lines up properly and everything looks as it would in a terminal window; the display of the code box when reading the post is restricted to a maximum width/length with scroll bars (tablet users can swipe within the widow) - this means that a casual reader with no need to view the detail of the code box text can easily scroll past it and not have a huge length of normal post text to find the end of.

For example (the code is much longer than the window it is displayed in):

Code: 
60 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/3 The Spy Who Came In From The Cold (2 of 3).ts - Already decrypted.
59 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/4 The Looking Glass War (2 of 2).ts - Already decrypted.
58 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/Smiley's People_20150127_2202.ts - Already decrypted.
57 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/8 The Secret Pilgrim (3 of 3).ts - Already decrypted.
56 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/4 The Looking Glass War (1 of 2).ts - Already decrypted.
55 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/2 A Murder of Quality (1 of 1).ts - Already decrypted.
54 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/7 Smiley's People (3 of 3).ts - Already decrypted.
53 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/8 The Secret Pilgrim (2 of 3).ts - Already decrypted.
52 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/6 The Honourable Schoolboy (3 of 3).ts - Already decrypted.
51 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/7 Smiley's People (2 of 3).ts - Already decrypted.
50 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/5 Tinker, Tailor, Soldier, Spy (3 of 3).ts - Already decrypted.
49 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/5 Tinker, Tailor, Soldier, Spy (1 of 3).ts - Already decrypted.
48 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/7 Smiley's People (1 of 3).ts - Already decrypted.
47 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/5 Tinker, Tailor, Soldier, Spy (2 of 3).ts - Already decrypted.
46 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/6 The Honourable Schoolboy (2 of 3).ts - Already decrypted.
45 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/Smiley's People_20150203_2200.ts - Already decrypted.
44 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/3 The Spy Who Came In From The Cold (3 of 3).ts - Already decrypted.
43 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/Smiley's People_20150106_2208.ts - Already decrypted.
42 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/Smiley's People_20150210_2201.ts - Already decrypted.
41 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/6 The Honourable Schoolboy (1 of 3).ts - Already decrypted.
40 25/09/2016 19:05:17 - /media/My Video/The Complete Smiley/8 The Secret Pilgrim (1 of 3).ts - Already decrypted.
39 25/09/2016 19:05:17 - DECRYPT: [/media/My Video/The Complete Smiley]
38 25/09/2016 19:05:17 - [/media/My Video/The Complete Smiley]
37 25/09/2016 19:05:17 - [/media/My Video/[ModSettings]/nfs/*Then edit options in folder]
36 25/09/2016 19:05:17 - [/media/My Video/[ModSettings]/nfs/*Create folder with mount name (opt+)]
35 25/09/2016 19:05:17 - [/media/My Video/[ModSettings]/nfs]
34 25/09/2016 19:05:17 - [/media/My Video/[ModSettings]/smb/HDRFOX3/host=192_168_1_13]
33 25/09/2016 19:05:17 - [/media/My Video/[ModSettings]/smb/HDRFOX3/mac=ABABABABABAB (only needed for wakeUp)]
32 25/09/2016 19:05:17 - [/media/My Video/[ModSettings]/smb/HDRFOX3/domain=WORKGROUP]
31 25/09/2016 19:05:17 - [/media/My Video/[ModSettings]/smb/HDRFOX3/wakeConstantly?]
30 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX3/password=Password]
29 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX3/shareFolder=off]
28 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX3/user=User]
27 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX3/folder=My Video]
26 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX3/wakeNow?]
25 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX3]
24 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/*Then edit options in folder]
23 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4/mac=ABABABABABAB (only needed for wakeUp)]
22 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4/wakeConstantly?]
21 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4/host=192_168_1_14]
20 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4/password=Password]
19 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4/shareFolder=off]
18 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4/user=User]
17 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4/folder=My Video]
16 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4/wakeNow?]
15 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4/domain=Domain or Workgroup]
14 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/HDRFOX4]
13 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb/*Create folder with mount name (opt+)]
12 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]/smb]
11 25/09/2016 19:05:16 - Special folder, overriding recursion.
10 25/09/2016 19:05:16 - [/media/My Video/[ModSettings]]
9 25/09/2016 19:05:16 - /media/My Video/Family Guy does Star Wars/Family Guy_20130718_2304.ts - Already decrypted.
8 25/09/2016 19:05:16 - /media/My Video/Family Guy does Star Wars/Family Guy_20130717_2302.ts - Already decrypted.
7 25/09/2016 19:05:16 - /media/My Video/Family Guy does Star Wars/Family Guy_20130716_2303.ts - Already decrypted.
6 25/09/2016 19:05:16 - DECRYPT: [/media/My Video/Family Guy does Star Wars]
5 25/09/2016 19:05:16 - [/media/My Video/Family Guy does Star Wars]
4 25/09/2016 19:05:16 - DECRYPT: [/media/My Video/Auto-Shrink]
3 25/09/2016 19:05:16 - [/media/My Video/Auto-Shrink]
2 25/09/2016 19:05:16 - /media/My Video/Weatherview/Weather for the Week Ahead_20160921_0030.ts - Already decrypted.
1 25/09/2016 19:05:16 - /media/My Video/Weatherview/Weather for the Week Ahead_20160923_0031.ts - Already decrypted.

This is achieved by selecting "code" ( </> symbol) from the post editor tool bar (possibly on a drop-down menu that looks like a three dots - meaning "more options") and pasting the text into the dialogue box that appears, or create a code box directly by enclosing the text in [code]...[/code] tags. For info see Help >> BB Codes (bottom right corner of forum page) or http://hummy.tv/forum/help/bb-codes.

Note the </> tool is not the same as the >_ tool. The latter is [code]...[/code], meaning "inline code", used for inserting a short snippet of code within a line of text, and does not confer the same ability to "viewport" a listing.

As mentioned above, you can go back and edit your posts to correct errors of presentation - so there's no excuse.
Click to expand...

SunnysideUK said:
n.b. This site was inaccessible for a while recently.
Click to expand...
See https://hummy.tv/forum/threads/forum-software-upgrade-planned-for-20-11-2024.11444/post-176769. If you kept an eye on recent posts, you would know when there's something afoot (and where to report problems).

SunnysideUK said:
If it is what's the format command to use please.
Click to expand...
Menu >> Settings >> System >> Data Storage >> Format Storage

SunnysideUK said:
and does it only effect the recording, not settings?
Click to expand...
It does not affect settings, but you will have to reload CF packages.
 
No, it's very wise. The Humax formatter is cr@p. You are better off doing it manually.
This has been posted several times before - see here.
You may need /dev/sdb instead of /dev/sda if you have the virtual-disk2 package installed.
This of course needs to be done from maintenance mode having run umount /dev/sda2 and similar first.
 
prpr said:
No, it's very wise. The Humax formatter is cr@p. You are better off doing it manually.
This has been posted several times before - see here.
You may need /dev/sdb instead of /dev/sda if you have the virtual-disk2 package installed.
This of course needs to be done from maintenance mode having run umount /dev/sda2 and similar first.
Click to expand...
Although fixdisk runs for some time and eventually aborts, I have still managed to read of 1.7TiB of recordings via the VirtualBox and Linux Mint on my Win7 laptop (although I found the bus of the 2-bay powered USB3 caddy to hang many times during copy, necessitating a USB reconnect. I don't know what the causes of this is).
After this I dismounted the partition with the disks GUI utility and used the virtual terminal to run the format:

sudo mkfs.ext3 -M 0 -O spare -T largefile /dev/sdb2

After, I gave the volume name “hmx_int_stor” as is wash showing its UUID/GUID (I don’t know if is important but that was its label before re-formatting)”.
Then copied some recording over to the partition.
I put this in a Humax and was presented and the remote could play the files but the Webif presented its install process.
I removed the drive and copied over the /mod directory.
Then all was well, just did the package updates.

I’ve since added the rest of the recording without problem.
I have noticed that the files are owned not by root but my virtual login, but this doesn’t seem to cause a problem

Q's.
1.Is there any permissions/ownership I should change?
2.Should I run fixdisk on it as a check?

Thanks for your help.
 
SunnysideUK said:
sudo mkfs.ext3 -M 0 -O spare -T largefile /dev/sdb2
Click to expand...
That's "-m 0 -O sparse_super".
After, I gave the volume name “hmx_int_stor” as is wash showing its UUID/GUID (I don’t know if is important but that was its label before re-formatting)”.
Click to expand...
It seems not to be important, but I do it too.
I have noticed that the files are owned not by root but my virtual login, but this doesn’t seem to cause a problem
Click to expand...
That depends how you did the copies. It's best to do everything as root to preserve the attributes/ownership. I use "rsync -av ..." to do stuff like this.
 
prpr said:
That's "-m 0 -O sparse_super".

It seems not to be important, but I do it too.

That depends how you did the copies. It's best to do everything as root to preserve the attributes/ownership. I use "rsync -av ..." to do stuff like this.
Click to expand...
I actually used:
Code: 
sudo mkfs.ext3 -m 0 -O sparse_super -T largefile /dev/sdb2
I was copying from my laptop log file to my iPad. Typos and auto-correction=Chinese whispers anyone?
 
Black Hole said:
No particular need, all the Humax firmware and CF software run as root so have access to everything.


I would!
Click to expand...
Problems getting maintenance mode. New post here: Oh, the link icon is not working. Well have it raw!
S

Thread 'Problem getting Maintenance mode'

I''m trying to get maintenance mode on one of my Humax recorders with the 2TB pipeline drive attached (its been working fine). but I'm having problems with this function;
I used webif to select maintenance and reboot (also tried shutdown/startup from remote). initially it looks OK but at about 52sec countdown I hear the disk park and display goes off. Ring stays blue, but its off (normal standby I suppose).
I'll bring it out of standby and maintenance is announced. Here it may repeat the problem (then on next reboot it will exit maintenance) or show my boot LCN "Freeview".
Telnet...
 
You must log in or register to reply here.
Back
Top